Vascular Access in Children: The Importance of Microsurgery for Creation of Autologous Arteriovenous Fistulae

Radial-to-cephalic arteriovenous fistulas6 (Fig. 1, Fig. 2) 

The incision is longitudinal, half way between the cephalic vein and the radial artery in the wrist. The nerves must be carefully preserved. The vein is freed, a ligature being placed on the collateral branches avoiding electric coagulation that could damage the vascular trunk itself. Saline irrigation must be frequent in order to prevent drying, and vessels are only handled by the adventitia. After section of the vein above a ligature, a longitudinal posterior incision of approximately 10mm in length is made in the proximal vein. The anterior wall of the artery is exposed. A longitudinal arteriotomy is made with a disposable ophthalmologic scalpel and completed with scissors. If a few drops of blood appear they are rinsed with heparin saline and dried with small sponges. The anastomosis is constructed with four running sutures, by suturing first the proximal angle. The running suture is left loose at the beginning so as to leave the lumen open. Frequent use of the zoom facilitates this very precise part of the operation and the needle must be repositioned in the needle holder for each passage through a vessel. Both proximal running sutures (posterior and anterior) are stopped at the middle of the arteriotomy. The anastomosis is then completed, starting from the distal angle. The tourniquet is then released. If there is a leak, an additional suture is added. Patency is confirmed by Doppler ultrasound, as the murmur can be absent during the first postoperative hours, particularly in small children. It is very important to ascertain that the vein is not stenosed at the upper limit of the dissection.

View Large Image Download to PowerPoint

Fig. 1. Completed microsurgical radiocephalic anastomosis in an infant.

View Large Image Download to PowerPoint

Fig. 2. Functioning radiocephalic AV fistula in a 13 year old child, one month post-operatively.

Ulnar-to-basilic arteriovenous fistulas 

A sufficient length of the vein must be freed, as it is situated at a distance from the artery. It can be helpful to place the forearm in flexion at the elbow during dissection. The artery is approached by partial excision of the flexor carpi ulnaris. The interval before puncture is longer than with a radial-to-cephalic AVF.

Brachial-to-cephalic arteriovenous fistulas 

A transverse incision at the anterior aspect of the elbow exposes the medial branch of the cephalic vein, which must be sufficiently freed to allow it to come into proximity with the brachial artery without traction. Here again, using a pneumatic tourniquet avoids having to free the artery.

Brachial-to-basilic arteriovenous fistulas with subcutaneous transposition 

The basilic vein follows a route which rapidly becomes deep above the elbow, meaning that it is often undamaged. Its use for angioaccess necessitates surgical transposition, even in the thinnest children. Its transposition is much easier when it is performed during a second stage, the first consisting of a simple AVF at the elbow which results in dilatation of the vein and thickening of its walls. The initial AVF is constructed using the medial branch of the basilic vein at the elbow or, if necessary, on its lateral branch. Transposition performed 2 months later requires a longitudinal incision on the medial side of the upper arm. Collaterals of the veins are ligated and divided. The vein is tunnelled very superficially along a straight line. Reimplantation on the artery is performed as low as possible. An interval of 3 weeks must be left before puncturing.

Femoral-to femoral arteriovenous fistulas, with transposition 

This type of vascular access is rarely used. Superficial transposition of the femoral vein is usually in a straight line with an anastomosis on the distal superficial femoral artery.

Arteriovenous bridge-graft 

In this case a graft (biograft or PTFE) is implanted between an artery and a vein. It is located under the skin in such a way as to be accessible to puncture.

With microsurgery 

Our first experience for AVF in 32 children (Table 1) was reported in 1978 (Coulonges S, Thèse de Médecine, Université Pierre et Marie Curie Paris VI) and we published our first results for AVF microsurgery in children under 10kg in 1981.7

Table 1. Arteriovenous angioaccesses in children: major publications

		Bourquelot8	Sanabia10	Lumsden14	Bagolan11	Sheth15
	Year of publication	1990	1993	1994	1998	2002
	Country	France	Spain	USA	Italy	USA
	Microsurgery	yes	yes	no	yes	no
	Number of accesses	434	86	61	112	52
	Failing to mature AVF	10%	10%	30%	5%	33%
	AVF vs. Graft	93%	100%	24%	100%	46%

In 19908 we reported 380 children undergoing microsurgery for AVF: the percentage of autologous AVF versus grafts was 93% and the creation of a distal AVF was possible in 78% of the children. The immediate patency rate was 96% and the 24-month patency rate was 85% in distal radial-cephalic AVF, 72% in brachial-basilic AVF, 47% in brachial-cephalic AVF and only 5% in AV bridge-grafts (Fig. 3).

View Large Image Download to PowerPoint

Fig. 3. Angioaccesses in children – Long term patency (N=434).

These benefits of microsurgery for children have been emphasized by a Canadian group9 in 1984, by Spanish workers in 199310 and by an Italian team in 1998.11

In a transverse survey of the three paediatric nephrology departments in Paris on 1 February 2003, we observed that 70% of 33 ESRD children were being haemodialyzed via an autologous fistula, 24% via a jugular CVC, and 6% were on peritoneal dialysis. This compared favourably with the annual publication of the North American Pediatric Renal Transplant Cooperative (NAPRTC) Study in 199612 reporting that between 1 January 1992 and 16 January 1996 two-thirds of children and adolescents on dialysis were maintained on peritoneal dialysis, despite an overall peritonitis rate of one episode every 13 patient-months, the majority of haemodialysis accesses were external percutaneous catheters (Fig. 4), with the subclavian vein the most common site. More recently the publication of NAPRTC in 200313 reported that, between 1992 and 1998, 70 to 80% of children and 59% of adolescents on dialysis were maintained on peritoneal dialysis and that younger children received haemodialysis almost exclusively through percutaneous catheters, while 57% of children more than six years old were dialyzed with fistula or graft after six months on haemodialysis.

View Large Image Download to PowerPoint

Fig. 4. Percentages of central venous catheter vs. arteriovenous access in American and French children. (*: 2001 NAPRTCS report, **: Paris 01/02/2003)

In 2001, we retrospectively reviewed 69 AVF in a young (mean age 20 years) and difficult cohort of 64 non-renal chronic disease patients requiring a permanent angioaccess for repeated transfusions, perfusions, apheresis and drug injections for sickle cell anaemia (N=19), parenteral nutrition (N=16), cystic fibrosis (N=9), haemophilia (N=6), hypercholesterolemia (N=5), haemochromatosis (N=3), and miscellaneous conditions (N=6) (Fig. 5). Although there had been no previous venous preservation strategy, it was possible with microsurgery to create distal AVF in 68% of cases; and insertion of a graft was necessary in only 4%. Long-term patency rate was around 60% after 10 years (Fig. 6).

View Large Image Download to PowerPoint

Fig. 5. Arteriovenous fistula in non-ESRD patients – Different pathologies.

View Large Image Download to PowerPoint

Fig. 6. Long-term patency of autologous arteriovenous fistulas in 64 non-renal chronic disease young patients.

Without microsurgery 

Limited-quality arteriovenous fistulas are frequently observed (Table 1). Proximal fistulas and grafts are easier to construct in children but they have high complication rates and they will destroy proximal veins and therefore threaten the long-term survival of the patient.

In 1994, Lumdsen reported on 61 angioaccesses in children without microsurgery; mean age was 11 years, only 25% were simple AVF and 30% of them failed to mature.14 The mean functional patency was 6.2 months; 76% of the accesses were grafts, mostly in the upper arm or in the thigh, with a mean patency of 10 months.

In 2002, Sheth15 reported on 52 arteriovenous angioaccesses in 13-year-old (mean age) children. The percentage of autologous AVF was low (46%) and the percentage of primary failures (failed to mature) was high (33%).

Curiously, these two authors demonstrated no interest in microsurgery and made no reference to any of the previous publications concerning microsurgical creation of AVF in children. Furthermore, Sheth stated wrongly that ‘the literature contains very little data regarding the success of permanent vascular access in paediatric patients’. In fact, the three European surgical teams who used microsurgery for angioaccesses in children (Table 1) reported only 5 to 10% percent of fistulas that failed to mature and 0 to 14% percent of grafts versus AVF. These results compare favourably with those of the two American teams not using microsurgery; 30 and 33% of their AVF failed to mature and 76 to 54% were grafts.

In 2005, Ramage16 published the results of a 20-year retrospective study without the use of microsurgery. One hundred twenty two arteriovenous vascular access procedures were performed on children, with a median age at initial access formation of 12.0 years (range, 4 weeks to 21.9 years). The most common procedure was CVC insertion (182 procedures) and then AVF formation (107 procedures), with only 15 AVGs created. Median censored survival was 3.14 years (95% confidence interval, 1.22 to 5.06) for AVFs and 0.6 years (95% confidence interval, 0.20 to 1.00) for CVCs. Factors adversely affecting vascular access survival were younger age, trainee operator, presence of hypoalbuminemia, and type of access undertaken, with AVF better than CVC.